COMPUTED tomography (CT) has transformed medical imaging, offering high-resolution, three-dimensional visualisations of anatomical structures. Its widespread adoption has significantly enhanced diagnostic accuracy across multiple medical disciplines. However, the increasing reliance on CT scans has raised concerns regarding radiation exposure and its potential long-term health effects on patients. This has created an ongoing challenge: maintaining optimal image quality while minimising radiation dose.
In recent years, artificial intelligence (AI) has emerged as a promising tool to address this issue. Deep learning algorithms, a subset of AI, have shown considerable potential in various aspects of CT imaging, including image reconstruction, noise reduction, and automated image analysis. AI-driven approaches aim to enhance image quality while reducing radiation doses, potentially improving diagnostic precision and patient safety.
Several studies have explored AI’s application in CT imaging across different anatomical regions and clinical scenarios. For example, research by Hu et al. investigated the use of deep learning models for intracranial aneurysm detection in CT angiography. Zhang et al. examined AI-driven reconstruction algorithms for CT imaging of sports injuries. Other studies have assessed AI’s role in breast cancer radiotherapy planning and liver CT imaging. Despite these advancements, the overall effectiveness of AI in improving CT image quality and reducing radiation dose remains an area of ongoing investigation. Variability in study designs, AI algorithms, and outcome measures makes it difficult to draw definitive conclusions about AI’s broader impact on CT imaging practices.
A recent meta-analysis synthesised evidence from five clinical validation studies to evaluate AI’s role in CT image quality control and radiation protection. The findings suggest that AI-based interventions significantly improve CT image quality and efficiency in image analysis, with a potential trend toward radiation dose reduction. The observed improvement in image quality was statistically significant (mean difference 0.70, 95% CI 0.43-0.96; P<.001), suggesting broad applicability across various clinical applications.
While the meta-analysis indicated a trend toward reducing the CT dose index with AI-based interventions, the result was not statistically significant (mean difference 0.47, 95% CI -0.21 to 1.15; P=.18). This highlights the need for further research to determine AI’s effectiveness in reducing radiation dose across different CT applications. Nonetheless, the potential for improved image quality and efficiency in image analysis suggests that AI could play a critical role in the future of CT imaging, enhancing diagnostic accuracy while mitigating radiation exposure risks.
Katie Wright, EMJ
Reference
Zhang S et al. Effectiveness of AI for enhancing computed tomography image quality and radiation protection in radiology: systematic review and meta-analysis. J Med Internet Res. 2025;27:e66622.
